DGX Spark 128GB vs RTX PRO 6000 Blackwell for local LLMs
- DGX Spark 128GB: Uses its largest memory option, 128 GB, with 8k context and an f16 KV cache.
- RTX PRO 6000 Blackwell: Assumes 32 GB of DDR5-5600 system RAM, Windows with this GPU driving the display, 8k context and an f16 KV cache.
What to know before choosing
The RTX PRO 6000 Blackwell is a workstation card with 96GB of VRAM at 1,792 GB/s; this page uses the 600W Workstation Edition with 32GB of system RAM. The DGX Spark here is one 128GB unit, whose GPU can use 116.9 GB in CanRun's count. The 300W Max-Q has the same memory and bandwidth, so CanRun gives it the same generation estimates and a lower prompt-processing estimate on its own NVIDIA RTX PRO 6000 Blackwell Max-Q page. Llama 3.3 70B fits the card's VRAM and runs great there, and slowly on the DGX Spark, a public benchmark.
Big low-bit files sit near both memory limits. Solar Open 2 250B at Q2_K runs well on both: on the DGX Spark's GPU, and on the card with some experts in system RAM, a theoretical estimate, since it needs 95.7 GB, just over the 95.4 GB CanRun counts from the listed 96GB; the card reports a little more to llama.cpp, so it may fit whole. DeepSeek V4 Flash 0731 at Q2_K splits them: it does not fit one DGX Spark in CanRun's count and runs well on the RTX PRO 6000 with experts in RAM, also theoretical.
GLM-5.3-Flash at Q4_K_M fits neither one DGX Spark nor the card with 32GB of RAM. Two linked 128GB units run it well with llama.cpp, a theoretical estimate; see the NVIDIA DGX Spark page.
Speed
The RTX PRO 6000 Blackwell generates text about 5.9× faster than the DGX Spark 128GB (median of our estimates over the 56 models that fit entirely in memory on both at Q4_K_M or the closest available quant).
- Reference model
- Ling-3.0 Flash at Q4_K_M, the largest model that fits entirely in memory on both at Q4_K_M or the closest available quant
- DGX Spark 128GB
- est. 52.8 tok/s (36.9–68.6, theoretical estimate ±30%)
- RTX PRO 6000 Blackwell
- est. 193.8 tok/s (135.7–252.0, theoretical estimate ±30%)
- Prompt processing on Ling-3.0 Flash
- DGX Spark 128GB: about 1,129 tok/s (678–1,581, rough estimate ±40%); RTX PRO 6000 Blackwell: about 6,742 tok/s (4,045–9,439, rough estimate ±40%)
Specs side by side
| Spec | DGX Spark 128GB | RTX PRO 6000 Blackwell | Difference |
|---|---|---|---|
| Memory | 128 GB of unified memory | 96 GB | — |
| Memory bandwidth | 273.0 GB/s | 1,792.0 GB/s | 6.56× on the RTX PRO 6000 Blackwell |
| FP16 compute | 62.5 TFLOPS | 126.0 TFLOPS | 2.02× on the RTX PRO 6000 Blackwell |
| Launch year | 2025 | 2025 | — |
| Price | $3,999 launch MSRP (128 GB) | — | — |
| Public benchmark results | 14 results | 1 result | — |
Verdicts at a glance
Out of 59 models at Q4_K_M (or the closest available quant) with 8k context.
- DGX Spark 128GBRuns great34 modelsRuns well12 modelsRuns slowly10 modelsWon't run3 models
- RTX PRO 6000 BlackwellRuns great56 modelsRuns well0 modelsRuns slowly0 modelsWon't run3 models
How to read the verdicts
- Runs great
- Fully on the GPU at 20 tok/s or more
- Runs well
- Fully on the GPU at 8–20 tok/s; MoE experts in system RAM at 20 tok/s or more; or at least 90% on the GPU at 8 tok/s or more
- Runs slowly
- 2–8 tok/s; CPU-only; less than 90% on the GPU; or MoE experts in system RAM below 20 tok/s
- Won't run
- Does not fit, or under 2 tok/s
Where the verdicts differ
22 models get a different verdict on the two.
Scroll sideways to see every column.
| Model | DGX Spark 128GB verdict | DGX Spark 128GB speed | RTX PRO 6000 Blackwell verdict | RTX PRO 6000 Blackwell speed |
|---|---|---|---|---|
| Qwen3.5 122B-A10BQ4_K_M | Runs well | 29.0 tok/s8k estimate 24.0 tok/s (19.2–28.8, calibrated estimate ±20%)measured (1 run, 4k context) | Runs great | est. 106.3 tok/s85.0–127.5calibrated estimate ±20% |
| Laguna S 2.1Q4_K_M | Runs well | est. 28.4 tok/s22.7–34.1calibrated estimate ±20% | Runs great | est. 112.6 tok/s90.1–135.1calibrated estimate ±20% |
| Solar Open 100BQ4_K_M | Runs well | est. 22.8 tok/s18.3–27.4calibrated estimate ±20% | Runs great | est. 102.3 tok/s81.9–122.8calibrated estimate ±20% |
| Llama 3.3 70BQ4_K_M | Runs slowly | 4.7 tok/s8k estimate 4.7 tok/s (3.8–5.7, calibrated estimate ±20%)measured (2 runs, 512 context) | Runs great | est. 27.7 tok/s24.4–31.1calibrated estimate ±12% |
| EXAONE 4.5 33BQ4_K_M | Runs slowly | est. 10.3 tok/s8.3–12.4calibrated estimate ±20% | Runs great | est. 60.9 tok/s53.6–68.2calibrated estimate ±12% |
| Nanbeige4.3 33BQ4_K_M | Runs slowly | est. 9.4 tok/s7.6–11.3calibrated estimate ±20% | Runs great | est. 55.7 tok/s49.0–62.3calibrated estimate ±12% |
| DeepSeek R1 Distill Qwen 32BQ4_K_M | Runs slowly | est. 9.7 tok/s7.7–11.6calibrated estimate ±20% | Runs great | est. 57.0 tok/s50.2–63.9calibrated estimate ±12% |
| Qwen3 32BQ4_K_M | Runs slowly | 10.7 tok/s8k estimate 9.7 tok/s (7.8–11.7, calibrated estimate ±20%)measured (2 runs, 512 context) | Runs great | est. 57.3 tok/s50.4–64.1calibrated estimate ±12% |
| EXAONE 4.0 32BQ4_K_M | Runs slowly | est. 10.7 tok/s8.6–12.9calibrated estimate ±20% | Runs great | est. 63.1 tok/s55.5–70.7calibrated estimate ±12% |
| Gemma 4 31BQ4_K_M | Runs slowly | 10.1 tok/s8k estimate 11.2 tok/s (9.0–13.5, calibrated estimate ±20%)measured (1 run, 512 context) | Runs great | est. 66.1 tok/s58.1–74.0calibrated estimate ±12% |
| Muse Glimmer 30BQ4_K_M | Runs well | est. 12.2 tok/s9.8–14.7calibrated estimate ±20% | Runs great | est. 72.0 tok/s63.4–80.7calibrated estimate ±12% |
| Granite 4.2 30BQ4_K_M | Runs slowly | est. 10.7 tok/s8.6–12.9calibrated estimate ±20% | Runs great | est. 63.1 tok/s55.6–70.7calibrated estimate ±12% |
| Qwen3.5 27BQ4_K_M | Runs slowly | 11.3 tok/s8k estimate 12.1 tok/s (9.7–14.5, calibrated estimate ±20%)measured (1 run, 4k context) | Runs great | est. 71.1 tok/s62.6–79.6calibrated estimate ±12% |
| Qwen3.6 27BQ4_K_M | Runs well | est. 12.3 tok/s9.8–14.7calibrated estimate ±20% | Runs great | est. 72.3 tok/s63.6–80.9calibrated estimate ±12% |
| Qwen3.8 27BQ4_K_M | Runs slowly | est. 11.8 tok/s9.5–14.2calibrated estimate ±20% | Runs great | est. 69.8 tok/s61.4–78.2calibrated estimate ±12% |
| Gemma 3 27BQ4_K_M | Runs well | est. 12.4 tok/s9.9–14.8calibrated estimate ±20% | Runs great | est. 72.8 tok/s64.1–81.6calibrated estimate ±12% |
| Mistral Small 3.2 24BQ4_K_M | Runs well | est. 13.6 tok/s10.9–16.3calibrated estimate ±20% | Runs great | est. 80.0 tok/s70.4–89.6calibrated estimate ±12% |
| Qwen3 14BQ4_K_M | Runs well | est. 20.6 tok/s16.5–24.7calibrated estimate ±20% | Runs great | est. 121.3 tok/s106.7–135.8calibrated estimate ±12% |
| DeepSeek R1 Distill Qwen 14BQ4_K_M | Runs well | est. 20.1 tok/s16.1–24.1calibrated estimate ±20% | Runs great | est. 118.3 tok/s104.1–132.5calibrated estimate ±12% |
| HyperCLOVA X SEED Think 14BQ4_K_M | Runs well | est. 20.9 tok/s14.6–27.2theoretical estimate ±30% | Runs great | est. 123.0 tok/s86.1–160.0theoretical estimate ±30% |
| Phi-4Q4_K_M | Runs well | est. 19.8 tok/s15.9–23.8calibrated estimate ±20% | Runs great | est. 116.9 tok/s102.9–131.0calibrated estimate ±12% |
| Gemma 4 12BQ4_K_M | Runs well | est. 29.4 tok/s23.5–35.2calibrated estimate ±20% | Runs great | est. 172.9 tok/s152.2–193.7calibrated estimate ±12% |
Every model on both
Scroll sideways to see every column.
| Model | DGX Spark 128GB verdict | DGX Spark 128GB speed | RTX PRO 6000 Blackwell verdict | RTX PRO 6000 Blackwell speed |
|---|---|---|---|---|
| Ling-3.0 FlashQ4_K_M | Runs great | est. 52.8 tok/s36.9–68.6theoretical estimate ±30% | Runs great | est. 193.8 tok/s135.7–252.0theoretical estimate ±30% |
| Qwen3.5 122B-A10BQ4_K_M | Runs well | 29.0 tok/s8k estimate 24.0 tok/s (19.2–28.8, calibrated estimate ±20%)measured (1 run, 4k context) | Runs great | est. 106.3 tok/s85.0–127.5calibrated estimate ±20% |
| Laguna S 2.1Q4_K_M | Runs well | est. 28.4 tok/s22.7–34.1calibrated estimate ±20% | Runs great | est. 112.6 tok/s90.1–135.1calibrated estimate ±20% |
| gpt-oss-120bMXFP4 | Runs great | 58.7 tok/s8k estimate 50.8 tok/s (40.6–60.9, calibrated estimate ±20%)measured (1 run, 2k context) | Runs great | est. 237.1 tok/s189.7–284.5calibrated estimate ±20% |
| Solar Open 100BQ4_K_M | Runs well | est. 22.8 tok/s18.3–27.4calibrated estimate ±20% | Runs great | est. 102.3 tok/s81.9–122.8calibrated estimate ±20% |
| Kolibri 1Q4_K_M | Runs great | est. 52.2 tok/s41.7–62.6calibrated estimate ±20% | Runs great | est. 197.0 tok/s157.6–236.4calibrated estimate ±20% |
| Llama 3.3 70BQ4_K_M | Runs slowly | 4.7 tok/s8k estimate 4.7 tok/s (3.8–5.7, calibrated estimate ±20%)measured (2 runs, 512 context) | Runs great | est. 27.7 tok/s24.4–31.1calibrated estimate ±12% |
| K2-Horizon MoVA 36B-A4BQ4_K_M | Runs great | est. 37.7 tok/s26.4–49.1theoretical estimate ±30% | Runs great | est. 149.6 tok/s104.7–194.5theoretical estimate ±30% |
| Qwen3.5 35B-A3BQ4_K_M | Runs great | est. 61.9 tok/s49.5–74.2calibrated estimate ±20% | Runs great | est. 213.4 tok/s170.7–256.0calibrated estimate ±20% |
| Qwen3.6 35B-A3BQ4_K_M | Runs great | est. 70.5 tok/s56.4–84.6calibrated estimate ±20% | Runs great | est. 229.8 tok/s183.9–275.8calibrated estimate ±20% |
| MiniCPM-V 4.7 35B-A3BQ4_K_M | Runs great | est. 76.0 tok/s60.8–91.2calibrated estimate ±20% | Runs great | est. 239.5 tok/s191.6–287.4calibrated estimate ±20% |
| EXAONE 4.5 33BQ4_K_M | Runs slowly | est. 10.3 tok/s8.3–12.4calibrated estimate ±20% | Runs great | est. 60.9 tok/s53.6–68.2calibrated estimate ±12% |
| Nanbeige4.3 33BQ4_K_M | Runs slowly | est. 9.4 tok/s7.6–11.3calibrated estimate ±20% | Runs great | est. 55.7 tok/s49.0–62.3calibrated estimate ±12% |
| DeepSeek R1 Distill Qwen 32BQ4_K_M | Runs slowly | est. 9.7 tok/s7.7–11.6calibrated estimate ±20% | Runs great | est. 57.0 tok/s50.2–63.9calibrated estimate ±12% |
| Qwen3 32BQ4_K_M | Runs slowly | 10.7 tok/s8k estimate 9.7 tok/s (7.8–11.7, calibrated estimate ±20%)measured (2 runs, 512 context) | Runs great | est. 57.3 tok/s50.4–64.1calibrated estimate ±12% |
| EXAONE 4.0 32BQ4_K_M | Runs slowly | est. 10.7 tok/s8.6–12.9calibrated estimate ±20% | Runs great | est. 63.1 tok/s55.5–70.7calibrated estimate ±12% |
| Nemotron 3.5 Lightning 30B-A3BQ4_K_M | Runs great | est. 69.8 tok/s48.8–90.7theoretical estimate ±30% | Runs great | est. 305.7 tok/s214.0–397.4theoretical estimate ±30% |
| Gemma 4 31BQ4_K_M | Runs slowly | 10.1 tok/s8k estimate 11.2 tok/s (9.0–13.5, calibrated estimate ±20%)measured (1 run, 512 context) | Runs great | est. 66.1 tok/s58.1–74.0calibrated estimate ±12% |
| Qwen3 30B-A3B (2507)Q4_K_M | Runs great | est. 60.7 tok/s48.6–72.9calibrated estimate ±20% | Runs great | est. 195.0 tok/s156.0–234.1calibrated estimate ±20% |
| Muse Glimmer 30BQ4_K_M | Runs well | est. 12.2 tok/s9.8–14.7calibrated estimate ±20% | Runs great | est. 72.0 tok/s63.4–80.7calibrated estimate ±12% |
| Granite 4.2 30BQ4_K_M | Runs slowly | est. 10.7 tok/s8.6–12.9calibrated estimate ±20% | Runs great | est. 63.1 tok/s55.6–70.7calibrated estimate ±12% |
| Qwen3.5 27BQ4_K_M | Runs slowly | 11.3 tok/s8k estimate 12.1 tok/s (9.7–14.5, calibrated estimate ±20%)measured (1 run, 4k context) | Runs great | est. 71.1 tok/s62.6–79.6calibrated estimate ±12% |
| Qwen3.6 27BQ4_K_M | Runs well | est. 12.3 tok/s9.8–14.7calibrated estimate ±20% | Runs great | est. 72.3 tok/s63.6–80.9calibrated estimate ±12% |
| Qwen3.8 27BQ4_K_M | Runs slowly | est. 11.8 tok/s9.5–14.2calibrated estimate ±20% | Runs great | est. 69.8 tok/s61.4–78.2calibrated estimate ±12% |
| Gemma 3 27BQ4_K_M | Runs well | est. 12.4 tok/s9.9–14.8calibrated estimate ±20% | Runs great | est. 72.8 tok/s64.1–81.6calibrated estimate ±12% |
| Gemma 4 26B-A4BQ4_K_M | Runs great | 47.0 tok/s8k estimate 51.5 tok/s (41.2–61.8, calibrated estimate ±20%)measured (1 run, 32k context) | Runs great | est. 210.4 tok/s168.3–252.4calibrated estimate ±20% |
| Mistral Small 3.2 24BQ4_K_M | Runs well | est. 13.6 tok/s10.9–16.3calibrated estimate ±20% | Runs great | est. 80.0 tok/s70.4–89.6calibrated estimate ±12% |
| gpt-oss-20bMXFP4 | Runs great | est. 71.7 tok/s57.4–86.0calibrated estimate ±20% | Runs great | est. 338.9 tok/s271.1–406.7calibrated estimate ±20% |
| Kanana 1.5 15.7B-A3BQ4_K_M | Runs great | est. 62.2 tok/s49.7–74.6calibrated estimate ±20% | Runs great | est. 233.4 tok/s186.7–280.1calibrated estimate ±20% |
| Qwen3 14BQ4_K_M | Runs well | est. 20.6 tok/s16.5–24.7calibrated estimate ±20% | Runs great | est. 121.3 tok/s106.7–135.8calibrated estimate ±12% |
| DeepSeek R1 Distill Qwen 14BQ4_K_M | Runs well | est. 20.1 tok/s16.1–24.1calibrated estimate ±20% | Runs great | est. 118.3 tok/s104.1–132.5calibrated estimate ±12% |
| HyperCLOVA X SEED Think 14BQ4_K_M | Runs well | est. 20.9 tok/s14.6–27.2theoretical estimate ±30% | Runs great | est. 123.0 tok/s86.1–160.0theoretical estimate ±30% |
| Phi-4Q4_K_M | Runs well | est. 19.8 tok/s15.9–23.8calibrated estimate ±20% | Runs great | est. 116.9 tok/s102.9–131.0calibrated estimate ±12% |
| Gemma 3 12BQ4_K_M | Runs great | est. 27.2 tok/s21.7–32.6calibrated estimate ±20% | Runs great | est. 160.1 tok/s140.9–179.3calibrated estimate ±12% |
| Mellum2.1 12B-A2.5B ThinkingQ4_K_M | Runs great | est. 104.4 tok/s83.5–125.2calibrated estimate ±20% | Runs great | est. 334.9 tok/s267.9–401.8calibrated estimate ±20% |
| Gemma 4 12BQ4_K_M | Runs well | est. 29.4 tok/s23.5–35.2calibrated estimate ±20% | Runs great | est. 172.9 tok/s152.2–193.7calibrated estimate ±12% |
| Qwen3.5 9BQ4_K_M | Runs great | 35.2 tok/s8k estimate 34.7 tok/s (27.8–41.6, calibrated estimate ±20%)measured (1 run, 4k context) | Runs great | est. 204.4 tok/s179.8–228.9calibrated estimate ±12% |
| K2-Horizon 7BQ4_K_M | Runs great | est. 31.3 tok/s25.1–37.6calibrated estimate ±20% | Runs great | est. 184.5 tok/s162.4–206.7calibrated estimate ±12% |
| Qwen3 8BQ4_K_M | Runs great | 43.9 tok/s8k estimate 34.1 tok/s (27.3–41.0, calibrated estimate ±20%)measured (1 run, 512 context) | Runs great | est. 201.1 tok/s177.0–225.2calibrated estimate ±12% |
| DeepSeek R1 0528 Qwen3 8BQ4_K_M | Runs great | est. 34.1 tok/s27.3–41.0calibrated estimate ±20% | Runs great | est. 201.1 tok/s177.0–225.2calibrated estimate ±12% |
| DeepSeek R1 Distill Llama 8BQ4_K_M | Runs great | est. 35.5 tok/s28.4–42.6calibrated estimate ±20% | Runs great | est. 209.3 tok/s184.2–234.4calibrated estimate ±12% |
| Kanana 1.5 8BQ4_K_M | Runs great | est. 35.5 tok/s28.4–42.6calibrated estimate ±20% | Runs great | est. 209.3 tok/s184.2–234.4calibrated estimate ±12% |
| Llama 3.1 8BQ4_K_M | Runs great | est. 35.5 tok/s28.4–42.6calibrated estimate ±20% | Runs great | est. 209.3 tok/s184.2–234.4calibrated estimate ±12% |
| Gemma 4 E4BQ4_K_M | Runs great | est. 61.5 tok/s43.1–80.0theoretical estimate ±30% | Runs great | est. 257.7 tok/s180.4–335.0theoretical estimate ±30% |
| Ling-3.0 TinyQ4_K_M | Runs great | est. 179.8 tok/s125.8–233.7theoretical estimate ±30% | Runs great | est. 485.1 tok/s339.6–630.7theoretical estimate ±30% |
| Gemma 4 E2BQ4_K_M | Runs great | est. 115.9 tok/s81.1–150.6theoretical estimate ±30% | Runs great | est. 386.7 tok/s270.7–502.7theoretical estimate ±30% |
| Qwen3.5 4BQ4_K_M | Runs great | est. 68.7 tok/s55.0–82.5calibrated estimate ±20% | Runs great | est. 404.8 tok/s356.3–453.4calibrated estimate ±12% |
| Nanbeige4.2 3BQ4_K_M | Runs great | est. 34.8 tok/s24.3–45.2theoretical estimate ±30% | Runs great | est. 204.8 tok/s143.3–266.2theoretical estimate ±30% |
| Spark-X2.5 4BQ4_K_M | Runs great | est. 73.4 tok/s58.7–88.1calibrated estimate ±20% | Runs great | est. 432.3 tok/s380.4–484.1calibrated estimate ±12% |
| Llama 3.2 3BQ4_K_M | Runs great | est. 72.0 tok/s57.6–86.3calibrated estimate ±20% | Runs great | est. 423.9 tok/s373.0–474.7calibrated estimate ±12% |
| LFM2.5 2.6BQ4_K_M | Runs great | est. 118.0 tok/s82.6–153.4theoretical estimate ±30% | Runs great | est. 695.3 tok/s486.7–903.8theoretical estimate ±30% |
| MiniCPM5 2BQ4_K_M | Runs great | est. 111.4 tok/s89.1–133.6calibrated estimate ±20% | Runs great | est. 656.0 tok/s577.2–734.7calibrated estimate ±12% |
| HyperCLOVA X SEED 1.5BQ4_K_M | Runs great | est. 117.3 tok/s93.8–140.8calibrated estimate ±20% | Runs great | est. 691.0 tok/s608.1–773.9calibrated estimate ±12% |
| EXAONE 4.0 1.2BQ4_K_M | Runs great | est. 162.1 tok/s129.7–194.6calibrated estimate ±20% | Runs great | est. 955.1 tok/s840.5–1,069.8calibrated estimate ±12% |
| Llama 3.2 1BQ4_K_M | Runs great | est. 197.5 tok/s158.0–236.9calibrated estimate ±20% | Runs great | est. 1,163.2 tok/s1,023.6–1,302.7calibrated estimate ±12% |
| MiniCPM5 1BQ4_K_M | Runs great | est. 238.9 tok/s191.1–286.7calibrated estimate ±20% | Runs great | est. 1,407.3 tok/s1,238.5–1,576.2calibrated estimate ±12% |
| GLM-5.3 FlashQ4_K_M | Won't run | — | Won't run | — |
| DeepSeek V4 Flash 0731MXFP4 | Won't run | — | Won't runTry Q2_K (heavy quality loss): Runs well | est. 44.2 tok/s30.9–57.4theoretical estimate ±30% |
| Solar Open 2 250BIQ4_XS | Won't runTry Q2_K (heavy quality loss): Runs well | est. 22.0 tok/s17.6–26.3calibrated estimate ±20% | Won't runTry Q2_K (heavy quality loss): Runs well | est. 83.0 tok/s58.1–107.9theoretical estimate ±30% |
Models with a thinking mode spend extra tokens thinking, so their speed thresholds are 1.5× stricter (30 / 12 / 3 tok/s). With thinking turned off, a model may do one grade better.
Public benchmarks
Measured results on the DGX Spark 128GB
| Model | Quant | Backend | Context | Prompt (tok/s) | Generation (tok/s) | Flags | Source | Measured |
|---|---|---|---|---|---|---|---|---|
| gpt-oss-120b | MXFP4 | llama.cpp | 2k | 2,444.0 | 58.7 | CUDA, no mmap | github.com | 2026-02-05 |
| Gemma 4 31B | Q4_K_M | llama.cpp | 512 | 687.0 | 10.1 | CUDA | github.com | 2026-04-14 |
| Gemma 4 26B-A4B | Q4_K_M | llama.cpp | 32k | — | 47.0 | CUDA, depth 32768 | github.com | 2026-06-11 |
| Qwen3.5 35B-A3B | Q8_0 | llama.cpp | 2k | 2,572.0 | 49.2 | CUDA | github.com | 2026-03-12 |
| Qwen3.5 9B | Q4_K_M | llama.cpp | 4k | — | 35.2 | — | github.com | 2026-08-13 |
| Qwen3.5 27B | Q4_K_M | llama.cpp | 4k | — | 11.3 | — | github.com | 2026-08-13 |
| Qwen3.5 122B-A10B | Q4_K_M | llama.cpp | 4k | — | 29.0 | — | github.com | 2026-08-13 |
| Llama 3.3 70B | Q4_K_M | llama.cpp | 512 | 284.0 | 4.7 | CUDA, Vulkan | github.com | 2025-11-22 |
| Llama 3.3 70B | Q4_K_M | llama.cpp | 512 | — | 4.6 | CUDA | github.com | 2026-08-12 |
| Llama 3.3 70B | Q8_0 | llama.cpp | 512 | — | 2.9 | CUDA | github.com | 2026-08-12 |
| Llama 3.3 70B | Q2_K | llama.cpp | 512 | — | 7.2 | CUDA | github.com | 2026-08-12 |
| Qwen3 32B | Q4_K_M | llama.cpp | 512 | 630.0 | 10.5 | CUDA, Vulkan | github.com | 2025-11-22 |
| Qwen3 32B | Q4_K_M | llama.cpp | 512 | 762.0 | 10.8 | CUDA | github.com | 2025-12-05 |
| Qwen3 8B | Q4_K_M | llama.cpp | 512 | 3,168.0 | 43.9 | CUDA | github.com | 2025-12-05 |
Measured results on the RTX PRO 6000 Blackwell
| Model | Quant | Backend | Context | Prompt (tok/s) | Generation (tok/s) | Flags | Source | Measured |
|---|---|---|---|---|---|---|---|---|
| llama-2-7b | Q4_0 | llama.cpp | 512 | 14,855.0 | 274.2 | — | github.com | 2025-09-05 |
Public benchmarks we calibrate against. Their conditions (context, backend, flags) can differ from the estimates above.
Frequently asked questions
Which is faster for local LLMs, the DGX Spark 128GB or the RTX PRO 6000 Blackwell?
The RTX PRO 6000 Blackwell generates text about 5.9× faster than the DGX Spark 128GB (median of our estimates over the 56 models that fit entirely in memory on both at Q4_K_M or the closest available quant). Estimated speed on Ling-3.0 Flash at Q4_K_M: the DGX Spark 128GB est. 52.8 tok/s (36.9–68.6, theoretical estimate ±30%); the RTX PRO 6000 Blackwell est. 193.8 tok/s (135.7–252.0, theoretical estimate ±30%).
What is the largest model each can run entirely in memory?
The DGX Spark 128GB fits Ling-3.0 Flash at Q4_K_M (77.01 GB file) entirely in 128 GB of unified memory, at est. 52.8 tok/s (36.9–68.6, theoretical estimate ±30%). The RTX PRO 6000 Blackwell fits Ling-3.0 Flash at Q4_K_M (77.01 GB file) entirely in 96 GB, at est. 193.8 tok/s (135.7–252.0, theoretical estimate ±30%).
How many local LLMs run well on each?
At Q4_K_M with 8k context, out of 59 tracked models: the DGX Spark 128GB — 34 Runs great, 12 Runs well, 10 Runs slowly, 3 Won't run; the RTX PRO 6000 Blackwell — 56 Runs great, 0 Runs well, 0 Runs slowly, 3 Won't run. 22 models get a different verdict.
Speeds are estimates from memory bandwidth, calibrated against public benchmarks, and each one comes with an error band and a confidence label. Real results vary with drivers, backend, context length and thermals.